2. Scalability

"Scale is not a feature you can bolt on later."

What is Scalability?

The ability of a system to handle increased load by adding resources — without architectural changes.

Scaling Dimensions

Vertical Scaling (Scale Up)

• Bigger machines (more CPU, RAM, disk)
• Simple but has hard limits
• Single point of failure remains

Horizontal Scaling (Scale Out)

• More machines
• Theoretically unlimited
• Requires stateless design or distributed state

Key Concepts

Load Characteristics

• Read-heavy — Caching, read replicas, CDN
• Write-heavy — Sharding, write-ahead logs, eventual consistency
• Compute-heavy — Horizontal pods, job queues, batch processing
• Storage-heavy — Tiered storage, data lifecycle, compression

Capacity Planning

Capacity = Current load × Growth factor × Safety margin

• Know your limits before you hit them
• Load test to find breaking points
• Plan for 2x headroom minimum

Auto-Scaling

• Reactive — Scale based on current metrics (CPU, memory, queue depth)
• Predictive — Scale based on forecasted load
• Scheduled — Scale based on known patterns (business hours, events)

Topics

• ☐ Stateless vs. stateful services
• ☐ Database scaling patterns
• ☐ Caching strategies
• ☐ Queue-based load leveling
• ☐ Sharding strategies
• ☐ Consistent hashing
• ☐ Read replicas
• ☐ Connection pooling
• ☐ Rate limiting
• ☐ Backpressure

Patterns

Stateless Services

Request → Load Balancer → Any Instance → Shared State (DB/Cache)

Database Scaling

Caching Layers

Client → CDN → App Cache → Database Cache → Database

Anti-Patterns

• Distributed monolith
• Shared mutable state
• Synchronous chains
• N+1 queries
• Unbounded growth (logs, caches, queues)

Tools

Metrics to Watch

• Request latency (p50, p95, p99)
• Throughput (RPS)
• Queue depth
• Resource utilization (CPU, memory, connections)
• Error rates during scale events

Reading

• Google SRE Book: Chapter 22 (Addressing Cascading Failures)
• Designing Data-Intensive Applications: Chapters 1, 5-6

Regulatory and control mappings

• ISO 27001 Annex A.8 Technological Controls A.8.6 Capacity management. A.8.14 Redundancy of information processing facilities.
• DORA ICT Risk Management Art 11 capacity considerations within continuity.
• ITIL 4 Practices Capacity and Performance Management.

Atoms

• Horizontal vs Vertical Scaling

People-substrate cross-cluster

Scaling a team has the same shape as scaling a system: horizontal scaling needs delegation (secure attachment); vertical scaling caps at the operator's own capacity. CPTSD adaptations cap at vertical.

• Bridge essay: Interview Training as Applied Clinical Psychology -- Item 8 (over-indexing as the cap on scaling)
• Psychology: Caretaker syndrome -- the leader who doesn't scale; over-functioning at small N becomes the bottleneck at larger N
• Psychology: Hyper-independence -- the anti-pattern of "scale by adding more of me"; refuses the delegation horizontal scaling requires
• Psychology: Ownership psychology -- calibrated ownership is the substrate horizontal scaling depends on
• Developmental-position: Three-Lane Developmental Model -- Lane-1 substrate doesn't scale (vertical-only); Lane-2 substrate scales horizontally; the scaling property is a developmental property, not a willpower property
• Competency: Solutions Oriented (Leader) -- develops problem-solvers, not problem-reporters; the scale-out version of the competency